This is the last of fourteen posts I have done on Regular Expressions for Google Analytics. Now that I have learned them (and hopefully explained them), it’s time to have that introductory post (I always do like to work backwards.)

Here’s the reason. People skip the introductions to books, they don’t read manuals, they just want to figure out how to make “it” work, whatever it is.

Only once it works are we ready to say, OK, so what? Why do I care, what’s it good for, what’s it bad for?

What are Regular Expressions (RegEx)? They use characters on your keyboard (like * and ^), enabling you to create an expression that may or may not match a target expression. They have a strict set of rules — just like a programming language would — and it’s easy to make mistakes with them. (This is why I am a big user of this RegEx checking tool. ) So you will always have at least two expressions, the Regular Expression with the funny characters and the expression you are matching on your site, or in someone’s address, or keyword. Here’s a quick example of the Regular Expression vs. target expression issue: I can create a regular expression like this luna|robb?in and then match it against the keywords people used to come to my site to filter out all the times people used my company name or my own name, whether they spelled it right now not. In this example, the keywords were my target expressions. (Need to understand that pipe symbol in the RegEx? Need to understand the question mark in the RegEx?)

So why use them? The first reason that RegEx are worth caring about — if you use Google Analytics — is that Google cares. There are certain tasks Google just won’t let you do correctly without using RegEx. Examples that come quickly to mind are: take yourself out of the data using an IP address, creating a custom filter, creating a filter that enables you to see both your subdomain and your domain in the same profile. (The latter is just an example of the second example, a custom filter, but I mention it because you can read about it in the GA help section.)

Other great examples of custom filters: Create a filter to learn what words people actually type in to Google before they click on your AdWord, instead of just learning which AdWord gets credit. (I use this one all the time. The only hack I like better is this one.) Force all your reports to give you pages by title instead of URL.

OK, so we understand that they are needed for filters. But how about goals? After all, you can do a head match or an exact match, why go to the trouble of using a RegEx match?

One reason would be if you have two pages that are essentially the same goal or the same place in the funnel. So, for example, let’s say that when the visitor reaches either of these two pages, www.mysite.com/folder3/thanks.html and www.mysite.com/thanksalot.html, he has really achieved the same goal. By using RegEx, you can make your goal page in Google Analytics /thanks and whenever someone reaches either of those pages, the same goal (G1, or G2, or whichever one you choose) is incremented. Then if you happen to care about which page actually matters the most, you can easily go to the Content Optimization > Goals and Funnel Verification > Goal verification to see which page mattered the most.

Of course, if you have other pages on your site that match /thanks, you have to get more specific with your RegEx. However, I never forget the lesson that a friend taught me: keep your RegEx as simple as possible.

Another reason to use RegEx is when you are lazy. After all, just because there are a ton of variations, who wants to create a ton of iterations? The example above shows how to combine two into one, but what if you had 15 variations? Example: you have to be sure that your company is not in the analytics. Your company owns all the IP addresses from 72.77.12.26 through 72.77.12.40, inclusive. You sure don’t want to create 15 filters. Instead, you can use a regular expression like this: ^72\.77\.12\.(2[6-9]|3[0-9]|40) — it will capture all fifteen IP addresses. (The little carat ^ at the beginning says, don’t match if there is something before the 72. The backslashes \ turn the special dots into plain dots. This 2[6-9] means, match 26, 27, 28 and 29. This 3[0-9] says, match 30, 31, etc through 39. 40 says, match 40. The pipes are OR signs. So at the end of the expression, you’ll be matching to 26-29, OR 30-39 OR 40.)

Well, that’s it on RegEx for now. Here are all the prior posts in the series:

Backslashes \
Dots .
Carats ^
Dollars signs $
Question marks ?
Pipes |
Parentheses ()
Square brackets []and dashes -
Plus signs +
Stars *
Regular Expressions for Google Analytics: Now let’s Practice
Bad Greed
RegEx and Good Greed
Intro to RegEx
{Braces}
Minimal Matching

Now that I am done with this series, I’ll go back and make sure that all the posts links to all the other posts consistently. Done! Done! All done!

Robbin
LunaMetrics

It is always hard to know how many options to give a user. Too many and you lose the conversion; too few and they can’t achieve their goal (and you lose the conversion.)

VistaPrint is a well-known Internet printing company, and is arguably best-known for their business cards. I clicked first on the product I wanted, “premium business cards” on the home page and landed here:

So this was the first moment of frustration. It didn’t matter whether I scrolled up or down, I couldn’t find any action buttons except the opportunity to upload my artwork or choose one of their templates. (And I really wanted them to tell me that they would walk me through a process.) But anyway, I uploaded my file (using the link that I have the red arrow pointing to) and was very frustrated to see that they had taken my vertical card, assumed that it was horizontal, and then yelled at me (electronically) because the aspect ratio of my image did not match the “chosen template.”

Eventually I realized that I had landed on tab 3 of a four tab wizard, as you can see below:

Finally, by making them back up (i.e. choosing Tab Two), I got to tell them that I wanted a vertical, not horizontal card:

While writing this post, I noticed that there *is* a business card wizard, right at the top of the home page. (I never saw it and it is a great example of ad blindness.) Even if I had, it is a “free business card” wizard, whereby you choose one of their templates and get their advertising. There truly is a way to do this right, just not a very intuitive way…

Robbin Steif
LunaMetrics

This week, Eric Peterson wrote a post about web analytics for beginners. As usual, he had lots of great advice, but I have to disagree with the “blog part” of his post. His advice was…

“Read some of the web analytics weblogs to keep up on current happenings. I personally like mine (go figure), but I also highly recommend Avinash Kaushik and Gary Angel.”

Every thing else he wrote made sense. (Take the WAA’s Web Analytics Course. Read the Web Analytics Forum. Buy and read Eric’s book, Web Analytics DeMystified.) But if you are in marketing, are suddenly in charge of analytics (Eric’s scenario) and are still working to understand the difference between visits and unique visitors, or log files vs. page tagging — do you really want to read Eric’s blog? On Eric’s blog, you get to learn advanced web analytics, like creating a metric for visitor engagement. In the same vein, do you really want to read Gary Angel’s blog? On Gary’s blog, you get to learn advanced web analytics, like worst practices in Business Intelligence. (Gary, I am sure I mangled that completely. Sorry.) They are both excellent resources but they sure aren’t the place you start when you need to know what a KPI is.

Avinash definitely has some great newbie stuff on his blog. Like his 10/90 rule post. Or his Tips for Small Business Web Analytics success. (And I really wanted to show one of his very first articles on visitor segmentation, but I can’t find it.)

Does anyone else do newbie stuff? I went through my feed reader looking for who else I could recommend. If the beginner is beginning with Google Analytics, Justin Cutroni is an awesome place to start — but not everyone uses GA. David Rhee, a web analyst at Gateway Computer, wrote a short piece for new analysts that really rocks (so get it on your list, but it’s not a full tutorial.)

Wouldn’t it be great if someone who loved analytics and understood them read all the blogs (there aren’t really all that many) and then posted a link and a snippet to every beginner post?

One last thing. There is another important beginner resource, the Web Analytics Association Training Day. It is really designed with new analysts in mind. (I am pretty sure of that.) It will be May 6 in San Francisco, the day before the Emetrics Summit, and I see that you can already sign up for that event (by itself or in addition to the Summit) on the Emetrics summit registration page.

Robbin Steif
LunaMetrics

I first met web analyst Ian Houston at Emetrics in Santa Barbara 2006. He already knew who I was (and had already helped me in his quiet, incredibly technically capable way) online. When he told me that he owned his own company, Visioactive, and that he didn’t have a website, I was surprised. “Yeah,” he said, “I should just put a blog up there.”

And he did, finally, today. So if you’re interested in cutting edge web analytics, check out his blog. If you want to see a great “results not found: (for conversion geeks like me), type something into his internal search engine that you probably won’t find.

Robbin Steif
LunaMetrics

After the Emetrics Summit in October, I wrote this post about Visual Sciences. Soon afterwards, Bob Chatham (who was then the CMO of VS and is now SVP - Education for purchaser Web Side Story) wrote replies to all of my comments. And there they sat for almost three months, in my inbox. (I have no excuses other than work and family — but better late than never.) My comments from the original post are in italics and Bob’s should be indented:

Yesterday at the eMetrics Summit, I saw a one-hour demo of Visual Sciences. The “demo-er” was Bob Chatham, their CMO. It is incredibly awesome.

[BC] Thank you, I have to agree – it’s why I came to the company. It took me three or four demos just to believe that it’s real

.
I walked into the presentation knowing that the product is black and probably visual. I walked out trying to figure out why, beyond budget, companies would choose a different solution if they could choose Visual Sciences. (I did come up with some, see the bottom of this post.)

[BC] Don’t worry, there’s a white-background version for “executives.” We just picked black based on UI research that shows that it’w easier to stare at a black background for long periods.

Even budget shouldn’t be an issue with a properly-sized system considered over a 3-year period. The TCO [total cost of ownership] usually works in our favor given the capabilities and expansion headroom. But you have to be willing to look beyond the short-term 1-year CPM rates that most ASPs charge.

I won’t speak much to the implementation or technology. Among other reasons, you can’t easily put them in a box. They aren’t a client-side solution (you pretty much have to have software), but they sure aren’t an old-fashioned logfile solution either. Instead, I want to speak just to the features/benefits.

[BC] Actually, we *are* a client-side solution…and a server-side one…and a hybrid of both, if you want. I’m assuming that you’re talking about data collection methods here. We can use just JavaScript tags like any other ASP solution, or combine tags with logfiles or server-side collection based on Visual Sensor. If you’re speaking about deployment/ownership models, there’s a lot of latitude here as well. While the majority of our clients choose licensed software deployed behind their firewalls (about 2/3 of the customer base) the rest are either using our managed services (they own the software/hardware, we run/operate it on our premises or theirs), or an ASP-like setup where they effectively pay a CPM charge. There’s no political statement here about one being better than the other, just an ability to accommodate a client’s data collection and deployment requirements.

First, they have developed the capability (a “visual sensor”) to capture data from all sorts of sources, not just the web. RFID, bar codes, text files, call center logs all qualify. So it is no longer just a web analytic package, it becomes a total analytic solution. I don’t think they like the phrase 360 degree view, but if that isn’t a 360, what is?

[BC] It’s really a combination of Visual Sensor (for real-time event collection at the source, e.g., a web server) and Visual Load (for importing log files, campaign codes/costs, lookup tables, etc.) that builds the 360-degree view. The customer has to supply the common key across the sources so that we can join them (we don’t provide data scrubbing or matching services, but others, like Group One Software, Axciom, or Experian do).

They have a proprietary database — and maybe that term is already too “boxy” — maybe data storage or “place to keep their data” would be more accurate. This seems to be their primary secret sauce, and enables them to relate data in any way you want. (They call this n-dimensional.) They aren’t working with a traditional relational database, so the number of fields aren’t constrained and it is no longer incredibly expensive to have someone crawl into the database and expand the fields (in fact, there is no crawling in.) I kind of dismissed this as Bob spoke - it’s an expensive solution so even if you pay for an extra 10 eVars from Omniture, you probably pay less money - but then, you might need that info now, not when the LiveSupport people get to it. In some ways, it is easier to compare it to Google Analytics, where you get one custom variable and no more, but the customer sets are so different that that becomes a senseless comparison (albeit easy to understand.)

[BC] I’m not sure I understand the Google Analytics comparison. With VS, you can have as many “custom variables” as you want (since the schema is completely open and tailorable, they’re effectively all custom…). You get a standard “web analytics” schema for W3C logs out of the box for Visual Site. As for expensive…well, it’w probably more expensive not to be able to answer the questions that you need to answer…;-)

It is amazingly easy to create whatever report you want. We were five people watching them demo, and we couldn’t come up with a single request for a report that Bob Chatham couldn’t create from scratch within 5-10 seconds.

[BC] Hey…if you think 5-10 seconds is fast, you should see how fast web analytics guru Eric T. Peterson can do it. I’m just the marketing guy.

{note from Robbin - I deleted all the screen shot chatter that was in this spot because none of my screen shots were real jpegs and will be just awful in a post}

It’s real time. Almost all the other page tagging solutions are real time, but unlike VS don’t change while you watch unless you hit reload and wait 60 seconds.
So why doesn’t everybody buy it? First, it is really expensive.

[BC] Let me know if you want to review a TCO [total cost of ownership] model for the product. We’re working on better comparisons to conventional solutions.

I don’t know exactly how expensive but I have a pretty good idea. So it’s not a solution you buy lightly, and you sure don’t buy it without a full-time analyst or a team of analysts or you won’t learn much. Plus, if you don’t have time or resources to learn everything from your standard high-end solution (SC, HBX, WebTrends etc), then it’s really not worth spending the money. And let’s face it (this is my usual soapbox) - most people aren’t getting as much as they should out of their free Google Analytics. In fact, most sites either don’t have analytics, don’t know that they have analytics, or don’t care (but I digress…)

[BC] Agreed about the need for analysts with *any* high-end solution. As to why folks aren’t getting everything from their “standard” high-end solutions, maybe it’s because they’re…standard. And then they give up in frustration

.
Second, I think that the company would be smart to sell the product on its merits and not sell it against the competition so much. Maybe that’s just the IBMer in me (I was taught to NEVER mention the competition by name), but I think their product, in the right situation, solves so many problems that they don’s need to have a negative campaign.

[BC] Aw, I’m just overly sensitive. We spent four years in stealth mode building the product, ignoring what other people thought. If we mention other vendors, it’s really to facilitate comprehension based on what people are more familiar with. Any suggestions on how to make the product concepts more accessible/believable to reduce the time that it takes to get it?

Thanks Bob for all that wisdom. Apologies for waiting so long to post this (and if any little gibberish marks from the email crept in, double apologies!)

Robbin Steif
LunaMetrics

This is my next to last post in this Regular Expression (RegEx) series. I have been thinking about this post for a long time and yesterday someone asked me a question (which finally got me to write this). She wrote that she had two pages that she wanted to roll into one Google Analytics goal. She created the Regular Expression for it, ran it through Epikone’s RegEx Coach, and it worked — but it wasn’t working in GA. (More on the Coach below.)

The two pages were:

subdomain.mysite.com/folder/subfolder/GoalThree.php
subdomain.mysite.com/folder/subfolder/GoalThreesome.php

She sent me a long, complicated expression which wasn’t working for her and asked my opinion.

This is absolutely a case of putting Good Greed to work for you, we will see in a minute. As I wrote in my last post, Regular Expressions are very greedy and they match everything unless you tell them not to. This is a very hard concept to wrap your head around — it means that, among other things, all the stuff before the expression and all the stuff after it gets matched to random things (unless you tell it not to. Or there is nothing to match to.)

Anyway, I wrote her back and said, why don’t you just write an expression like this:

/folder/subfolder/GoalThree

This assumes that she doesn’t have other GoalThreeVersions that will be incorrectly mixed in here. If, for example, she had another page, /folder/subfolder/GoalThreeCornered, that would qualify as a match too (because the RegEx matches everything it can, even if those characters aren’t in the Regular Expression.) Moving back to how simple her RegEx might be, she might even have been able to get away with a goal like this, depending on her site:

/GoalThree

This matches every expression that includes /GoalThree

Finally a word about the Epikone RegEx coach. I haven’t talked to Justin about this. But I am fairly sure that the coach is configured to check whether the phrase you type is a match to the RegEx you type, using the way GA interprets RegEx. That doesn’t mean that you necessarily come up with a valid goal, or an IP address that will actually filter anything. For example, you might use it to see if colou?r is a valid RegEx for color and for colour (it should be), but that doesn’t mean colou?r will necessarily work in your Google Analytics profile filters or goals. You really have to understand the context in which you are using the expression and what GA demands of you in addition to correctly configuring two expressions to match each other.

Backslashes \
Dots .
Carats ^
Dollars signs $
Question marks ?
Pipes |
Parentheses ()
Square brackets []and dashes -
Plus signs +
Stars *
Regular Expressions for Google Analytics: Now let’s Practice
Bad Greed
RegEx and Good Greed
Intro to RegEx
{Braces}
Minimal Matching

Robbin
LunaMetrics

Blogger did an update yesterday (scheduled for the same day as an update of Google Groups.) My feeds are not working now for new subscribers (at least on my computer - I would love if someone else would try it and send me email or just comment) and I can’t even complain about it on the Blogger Google group… Another good reason to get off Blogger.

Sorry.

Robbin

I just couldn’t resist showing this tiny piece of a non-profit website.

Wouldn’t a check box next to “Please don’t show my name in your printed thank you material?” do a better job? Now that I’ve looked at it 15 times, it makes perfect sense but when I first read it, I felt like I had to parse all the negatives.

Robbin Steif
LunaMetrics

Three times in the past week, I have tried to make a purchase from a (big important secure) website, and every time, the site makes me feel like they couldn’t understand my goals.

The first site was iTunes. I wanted to buy a gift certificate, and so went right into my own iTunes setup and choose Give an iTunes gift from the iTunes store. But in the shopping cart, they made me sign in. Hmm, I thought, do they think I am buying something for myself? Are they going to charge my personal credit card, when I need to use a different card? And shouldn’t my grandmother be able to make this purchase without owning her own copy of iTunes? So I went to the Apple site and succeeded in purchasing a cerfticate there — but all they needed to say was, please sign in and you will get a chance to choose a credit card.

The second site was NetFlix. I wanted to buy a gift certificate (again), but they took me all the way to the screen where I had to sign in (like iTunes) and even told me that when I hit enter, the sale was final. However, they still hadn’t given me a chance to tell them where to mail the DVDs to, so obviously they were going to send them to me, right? I put this one off for two days and finally went back to it. Eventually I realized that the individual who gets the gift chooses when to redeem it and where to receive the DVDs, so I got back on my computer (as if I ever leave it) and finally made the purchase. (I also submitted the suggestion to them — after all, they are offering $1M for a better algorithm, maybe they are paying conversion scientists??)

The third one was PR-Web. I used to do a lot of work with them and for whatever reason, haven’t lately. Maybe their interface has changed and maybe I just don’t remember it, but once again, they made me do things in the “wrong” order without giving me any reassurance. In this case, they insisted on taking all my credit card information without telling me that “On the the next screen, you will have a chance to upload your press release.” I finally just did it in their backward manner because they do tell the customer that s/he can give them funds to sit there unused (like a no-interest checking account) — I figured I’d have a way to use the money even if I didn’t get a screen for the press release.

Just think - all I wanted was a little reassurance.

Robbin Steif
LunaMetrics

Much to my chagrin, I didn’t really discover Negative Exact Match. Example: -[keyword]

In fact, it is called embedded match. (From the Department of Non-Descriptive Names.) The example Google gives is in the entertainment industry — you sell movie merchandise but you don’t own a movie theater or sell DVDs, so you can’t sell the actual movie experience. So you wouldn’t want people to click through to your site (and spend your AdWords money) if they type in just the word Spiderman but you want all the variations, Spiderman tights and Spiderman costume etc.

Robbin Steif
LunaMetrics